Low birth weight and metabolic disease

It has long been recognised that there is a relationship between impaired early growth and the subsequent development of adult disease. Early epidemiological studies established the association between poor fetal growth and increased risk of developing CVD (Barker & Osmond, Reference Barker and Osmond1986), impaired glucose tolerance (Hales et al. Reference Hales, Barker, Clark, Cox, Fall, Osmond and Winter1991), and the metabolic syndrome (Barker et al. Reference Barker, Hales, Fall, Osmond, Phipps and Clark1993). The ‘thrifty phenotype hypothesis’ put forward by Hales and Barker in 1992 (see Hales & Barker, Reference Hales and Barker2001), sought to explain why offspring exposed to poor prenatal nutrition are more susceptible to disease in later life. They proposed that in response to impaired nutrient supply (either as a result of maternal malnutrition or reduced placental function), the growing fetus will make adaptations in utero in order to maximise metabolic efficiency in relation to the storage and usage of fuels, to increase chances of survival postnatally. The programming of such a phenotype will continue to be beneficial if conditions of poor nutrition are extended into postnatal life; however, in the presence of adequate or plentiful nutrition these adaptations become detrimental and predispose to the development of obesity and associated features of the metabolic syndrome. Over the past two decades further epidemiological evidence and results from animal models have provided clear support for this hypothesis (Hales & Barker, Reference Hales and Barker2001).

A reduction in fetal nutrition usually results in a retardation of growth and organ-specific alterations in development, including the preservation of brain growth at the expense of other peripheral organs, such as the kidneys and pancreas (Desai et al. Reference Desai, Crowther, Lucas and Hales1996). Results from experimental studies in animals have shown that in response to poor maternal nutrition there is a reduction in pancreatic cell mass (Snoeck et al. Reference Snoeck, Remacle, Reusens and Hoet1990) and in insulin secretory responses (Dahri et al. Reference Dahri, Snoeck, Reusens-Billen, Remacle and Hoet1991). Alterations in kidney development, including a reduced number of nephrons and associated glomerular hypertrophy, are seen in low-birth-weight subjects, and subsequently confer an increased risk of later renal failure (Zandi-Nejad et al. Reference Zandi-Nejad, Luyckx and Brenner2006). Changes in body composition, including an increase in adipose mass relative to lean mass, are also associated with low birth weight in man (Yajnik, Reference Yajnik2000; Loos et al. Reference Loos, Fagard, Beunen, Derom and Vlietinck2001). As will be addressed, the combination of these factors contributes a much greater risk of developing the serious chronic diseases type 2 diabetes and CVD.

Early growth, body composition and obesity development

Impaired fetal growth, leading to a low birth weight, is associated with increased risk of obesity in childhood and later adult life (Ravelli et al. Reference Ravelli, Stein and Susser1976; Gale et al. Reference Gale, Martyn, Kellingray, Eastell and Cooper2001; Kensara et al. Reference Kensara, Wootton, Phillips, Patel, Jackson and Elia2005). In addition to increasing obesity risks, a low birth weight is associated with a tendency to store this excess adipose tissue centrally, which is an additional predictor of later metabolic disturbance (Yajnik, Reference Yajnik2003; Bosy-Westphal et al. Reference Bosy-Westphal, Geisler, Onur, Korth, Selberg, Schrezenmeir and Muller2005). At the other end of the spectrum, overnutrition during fetal and early life is also associated with a detrimental disease profile in later life (Armitage et al. Reference Armitage, Taylor and Poston2005). Both excessive maternal food intake and the consumption of a high-fat diet are associated with increased risks of CVD (Newman et al. Reference Newman, Freedman, Voors, Gard, Srinivasan, Cresanta, Williamson, Webber and Berenson1986) and impaired glucose tolerance and diabetes development (McCance et al. Reference McCance, Pettitt, Hanson, Jacobsson, Bennett and Knowler1994). Thus, the relationship between birth weight and metabolic disease is such that both those born small or large show increased risk of obesity in later life, leading to a ‘J’- or ‘U’-shaped relationship (Parsons et al. Reference Parsons, Power and Manor2001).

Growth during pregnancy, and also during the early postnatal period, plays an important role in long-term health. Individuals who are born small but subsequently undergo catch-up growth during the early postnatal period have a higher extent of adiposity in later childhood and adult life, and increased insulin resistance (Leon et al. Reference Leon, Koupilova, Lithell, Berglund, Mohsen, Vagero, Lithell and McKeigue1996; Forsen et al. Reference Forsen, Eriksson, Tuomilehto, Reunanen, Osmond and Barker2000; Yajnik, Reference Yajnik2000; Jaquet et al. Reference Jaquet, Deghmoun, Chevenne, Collin, Czernichow and Levy-Marchal2005). In terms of glucose tolerance, reduced insulin sensitivity is predicted by increased weight gain during childhood (Yajnik, Reference Yajnik2000; Ong & Dunger, Reference Ong and Dunger2004). In a study of low-birth-weight individuals the greatest extent of glucose intolerance was found to be present in those who were born small and were currently obese (Barker et al. Reference Barker, Hales, Fall, Osmond, Phipps and Clark1993). The findings of these epidemiological studies have been confirmed using animal models, with low birth weight and subsequent development of obesity contributing independently and in an additive manner (Petry et al. Reference Petry, Ozanne, Wang and Hales1997).

As well as alterations in fat mass, low-birth-weight individuals attain a lower lean muscle mass and strength and reduced bone density (Gale et al. Reference Gale, Martyn, Kellingray, Eastell and Cooper2001; Sayer & Cooper, Reference Sayer and Cooper2005). Given that muscle is the primary tissue involved in the uptake of glucose, a reduction in the amount of muscle mass would also predict a reduced glucose tolerance, leading to a more rapid or severe decline in metabolic control. In addition, it could also be predicted that a reduction in muscle mass itself would impact on an individual's ability to initiate exercise.

The maternal nutritional and endocrine environment, clearly, play a key role in the development of a growing fetus. Offspring of mothers with diabetes or impaired glucose tolerance, and who are therefore exposed to elevated levels of insulin, glucose and leptin, have increased risks of becoming obese (Plagemann et al. Reference Plagemann, Harder, Franke and Kohlhoff2002; Catalano et al. Reference Catalano, Kirwan, Haugel-de Mouzon and King2003). In addition, infants of diabetic mothers who are fed breast milk from normoglycaemic women have a reduced risk of obesity in later childhood compared with those fed maternal milk (Plagemann et al. Reference Plagemann, Harder, Franke and Kohlhoff2002). Breast-feeding itself has been found to lower obesity risk compared with formula feeding (Dewey, Reference Dewey2003). Postulated reasons for this association are the typically reduced energy intakes and growth rates associated with breast-fed infants (Heinig et al. Reference Heinig, Nommsen, Peerson, Lonnerdal and Dewey1993), and altered satiety responses or programmed differences in appetite regulation may also play a role.

Despite the abundant epidemiological evidence for the role of poor early growth predisposing to increased adipose accretion in later life, and its support in animal models, there is a relative paucity of information regarding the underlying mechanisms. There is clearly a need for further research in this area in order to determine the mechanisms involved in this response. In terms of programming effects contributing towards enhanced catch-up fat, it would be expected that in response to undernutrition in utero, adipocytes may have an enhanced ability to store fat. However, a recent study (Bieswal et al. Reference Bieswal, Hay, McKinnon, Reusens, Cuignet, Rees and Remacle2004) has failed to find differences in either the proliferation or lipid-accumulating properties in vitro of adipocytes isolated from growth-retarded rats compared with controls. There is some evidence that offspring of protein-restricted dams show an increased hepatic expression of lipogenic enzymes (Maloney et al. Reference Maloney, Gosby, Phuyal, Denyer, Bryson and Caterson2003); however, whether permanent alterations in fuel partitioning may result following early developmental insult remains to be fully investigated.

Regulation of catch-up growth

Increasingly, the role of catch-up growth is receiving attention as being key in the development of metabolic disease. In addition, the fact that tissue-specific alterations occur is now providing some clues as to how the body responds to a period of deprivation to maximise fuel storage in case of repeated future nutritional restriction. Dulloo and colleagues (see Cettour-Rose et al. Reference Cettour-Rose, Samec, Russell, Summermatter, Mainieri, Carrillo-Theander, Montani, Seydoux, Rohner-Jeanrenaud and Dulloo2005, Dulloo, Reference Dulloo2006) in a number of studies have shown that suppressed thermogenesis is a central process in catch-up growth. More specifically, it has been shown (Cettour-Rose et al. Reference Cettour-Rose, Samec, Russell, Summermatter, Mainieri, Carrillo-Theander, Montani, Seydoux, Rohner-Jeanrenaud and Dulloo2005) that during periods of re-feeding following semi-starvation, a state that induces suppression of thermogenesis in proportion to fat-store depletion, there is a redistribution of glucose such that skeletal muscle uptake is reduced while adipose tissue uptake is enhanced. Furthermore, adipose tissue fatty acid synthase activity is increased in these re-fed animals. Thus, it appears that coordinated alterations in tissue glucose handling confer a thrifty phenotype and that tissue-specific alterations in insulin sensitivity are involved. The way in which the body senses the reduction in adipose stores, and the signals and effector systems that mediate these responses, remain largely obscure (Dulloo, Reference Dulloo2006).

Hypothalamic regulation of energy balance

In normal individuals the regulation of energy balance involves a matching of energy intake with expenditure in order to maintain a relatively stable body weight over an extended period of time. This process should be simple enough, yet in a modern Westernised society, with an overabundance of readily-available foods and an increasingly sedentary lifestyle, excess energy deposition, in the form of adipose tissue, challenges this regulation. When peripheral energy stores are sufficient circulating factors are released in proportion to these stores and they target key centres in the brain to coordinate a reduction in feeding, an increase in expenditure and a suppression of endogenous glucose production (Schwartz et al. Reference Schwartz, Woods, Porte, Seeley and Baskin2000). Insulin was the first of such satiety signals to be identified. Following feeding, circulating insulin levels increase and insulin enters the brain via a regulated transport mechanism (Baura et al. Reference Baura, Foster, Porte, Kahn, Bergman, Cobelli and Schwartz1993) where it acts to reduce energy intake and body weight (Woods et al. Reference Woods, Lotter, McKay and Porte1979). The second key hormone that has been shown to act as such a signal of plentiful fat stores is leptin, a peptide hormone that is secreted predominantly by adipocytes. In the adult circulating leptin levels are related to body mass (Maffei et al. Reference Maffei, Halaas, Ravussin, Pratley, Lee and Zhang1995), but they are more closely correlated with the extent of adiposity (Liuzzi et al. Reference Liuzzi, Savia, Tagliaferri, Lucantoni, Berselli, Petroni, De Medici and Viberti1999). Leptin, like insulin, is also transported across the blood–brain barrier by a saturable system (Banks et al. Reference Banks, Kastin, Huang, Jaspan and Maness1996).

Numerous early ablation studies (Schwartz et al. Reference Schwartz, Woods, Porte, Seeley and Baskin2000) have indicated a key role for the hypothalamus in the regulation of energy balance. Within the hypothalamus the arcuate nucleus (ARC) is considered to be the initial site to receive input from circulating factors such as insulin and leptin. Two key populations of neurons in the ARC are involved in the ‘first-order’ sensing of energy stores. The first population express the orexigenic appetite-enhancing signals neuropeptide Y and agouti-related protein. The second population of neurons are those that express the anorexigenic pro-opiomelanocortin and cocaine- and amphetamine-regulated transcript neuropeptides. Centrally, an increase in insulin or leptin signalling acts to inhibit neuropeptide Y/agouti-related protein neurons of the ARC, while at the same time activating pro-opiomelanocortin/cocaine- and amphetamine-regulated transcript cells, leading ultimately to a reduction in feeding and increase in energy expenditure (Schwartz et al. Reference Schwartz, Woods, Porte, Seeley and Baskin2000). Conversely, in states of negative energy balance (such as starvation or fasting) leptin and insulin levels drop, feeding responses are increased and expenditure is reduced (Ahima, Reference Ahima2000). Thus, in addition to conveying information regarding whole-body adipose stores, a key role for leptin is in the signalling of more short-term changes in energy flux (Ahima & Flier, Reference Ahima and Flier2000). Rapid changes in plasma leptin levels during fasting, in the absence of any change in total adipose tissue mass, are key in mediating the neuroendocrine adaptation to a fasting state (Ahima et al. Reference Ahima, Prabakaran, Mantzoros, Qu, Lowell, Maratos-Flier and Flier1996).

Appropriate peripheral secretion and central sensing of leptin and insulin have recently been recognised as key in the maintenance of body weight and glucose homeostasis (Schwartz & Porte, Reference Schwartz and Porte2005). As in peripheral tissues, hypothalamic insulin signalling involves the insulin receptor substrate–1-phosphatidylinositol 3-kinase pathway (Niswender et al. Reference Niswender, Morrison, Clegg, Olson, Baskin, Myers, Seeley and Schwartz2003). It is also now clear that leptin and insulin share a common signalling pathway at the level of the hypothalamus (Schwartz & Niswender, Reference Schwartz and Niswender2004). Leptin receptor activation, which leads to the induction of suppressor of cytokine signalling-3 and prevents further signalling at the level of the receptor, also negatively influences insulin signalling (Ueki et al. Reference Ueki, Kondo, Tseng and Kahn2004). In obese subjects both insulin and leptin levels are elevated (Liuzzi et al. Reference Liuzzi, Savia, Tagliaferri, Lucantoni, Berselli, Petroni, De Medici and Viberti1999), but these individuals fail to respond appropriately. Central resistance to the anorectic and metabolic effects of these hormones is therefore thought to underlie many of the metabolic complications that arise during obesity. Interestingly, however, it appears that chronic hyperleptinaemia results in the development of selective leptin resistance. Although there is a loss of the appetite-suppressing and metabolism-enhancing effects of leptin, studies in rodents (Correia et al. Reference Correia, Haynes, Rahmouni, Morgan, Sivitz and Mark2002; Rahmouni et al. Reference Rahmouni, Correia, Haynes and Mark2005) have shown that there is a preservation of the effects of this hormone on the cardiovascular system. The reasons for this disparity in leptin resistance are as yet unclear, but it provides a mechanism whereby obesity, with its associated hyperleptinaemic state, contributes to elevated blood pressure and associated cardiovascular complications.

In addition to central effects on appetite, leptin also plays a key role in the regulation of peripheral tissue metabolism. Leptin receptors are present in many non-neural tissues, and both direct and indirect (central nervous system-mediated) effects of leptin act to increase sympathetic nervous system activity (Haynes et al. Reference Haynes, Sivitz, Morgan, Walsh and Mark1997). As well as up regulating energy expenditure via central nervous system-mediated pathways, leptin acts to suppress lipogenesis and promote lipid oxidation (Minokoshi et al. Reference Minokoshi, Kim, Peroni, Fryer, Muller, Carling and Kahn2002). This potent anti-steatotic role of leptin, opposing the fat-storing actions of insulin, is thought to be key in preventing ectopic deposition of lipid in non-adipose tissues (particularly muscle and pancreatic islets), and reducing lipotoxicity, a central process in the development of insulin resistance (Unger, Reference Unger2003).

Hypothalamic signalling: downstream of the arcuate nucleus

Relatively less understood is the circuitry downstream of the ARC signalling, and the way in which a coordinated alteration in feeding and energy expenditure is achieved. The pro-opiomelanocortin/cocaine- and amphetamine-regulated transcript and neuropeptide Y/agouti-related protein populations of the ARC, mentioned earlier, project to secondary sites within the hypothalamus, including the paraventricular nucleus, the lateral hypothalamic area, dorsal medial nucleus and the ventromedial nucleus. The melanocortin pathway is one efferent outflow that has received a large amount of attention in recent years, given its crucial role in the regulation of energy balance, both in human subjects and animal models. Indeed, in man genetic defects within the melanocortin pathway are the most common forms of monogenic obesity (O'Rahilly et al. Reference O'Rahilly, Farooqi, Yeo and Challis2003). The melanocortin peptides are derived from the cleavage of the pro-opiomelanocortin-precursor peptide, and they act through a family of receptors (MC1–5R) to inhibit food intake and increase energy expenditure (Cone, Reference Cone2005). In terms of energy balance the MC4R subtype is considered to be most crucial, which is supported by the following findings: mutations in this receptor result in severe early-onset obesity in human subjects (Yeo et al. Reference Yeo, Farooqi, Aminian, Halsall, Stanhope and O'Rahilly1998); disruption of this gene results in hyperphagia and development of obesity in mice (Huszar et al. Reference Huszar, Lynch, Fairchild-Huntress, Dunmore, Fang and Berkemeier1997); over-expression of the endogenous antagonist of the MC4R, agouti, also causes obesity (Klebig et al. Reference Klebig, Wilkinson, Geisler and Woychik1995).

In terms of the downstream integration and possible divergence of appetite and expenditure regulation, a recent study (Balthasar et al. Reference Balthasar, Dalgaard, Lee, Yu, Funahashi and Williams2005) has shown that there is a divergence in the control of these systems downstream of the ARC. Using the obese MC4R-knock-out mouse model MC4R expression was reactivated specifically in hypothalamic sites thought to be critical in mediating the melanocortin actions. When MC4R was restored in paraventricular nucleus and amygdala neurons, a 60% reduction in the obesity associated with the MC4R-knock-out phenotype was observed. Remarkably, this restoration was found to be associated with a correction of the hyperphagia in these animals, but had no effect on energy expenditure. Furthermore, there is now emerging evidence (Zhang & Scarpace, Reference Zhang and Scarpace2006) that targetting of the melanocortin pathway, downstream of leptin actions, may provide a means of circumventing leptin resistance and therefore be of potential clinical relevance in the treatment of obesity.

Leptin: a role in development of energy balance systems

A critical role for leptin in the regulation of body weight is illustrated by the obese phenotype of the ob/ob mouse model, which possesses a mutation in the leptin gene and is thus unable to secrete this hormone (Zhang et al. Reference Zhang, Proenca, Maffei, Barone, Leopold and Friedman1994). Administration of leptin to these animals restores body weight to normal as a result of both a suppression of feeding and an increase in thermogenesis (Campfield et al. Reference Campfield, Smith, Guisez, Devos and Burn1995; Halaas et al. Reference Halaas, Gajiwala, Maffei, Cohen, Chait, Rabinowitz, Lallone, Burley and Friedman1995; Pelleymounter et al. Reference Pelleymounter, Cullen, Baker, Hecht, Winters, Boone and Collins1995). Leptin administration to both mutant and wild-type animals promotes the mobilisation of fuels, produces an increase in O₂ consumption and favours the expenditure of lipid as opposed to carbohydrate stores, as shown by a reduced RQ (Hwa et al. Reference Hwa, Ghibaudi, Compton, Fawzi and Strader1996).

In addition to these metabolic effects in the adult, investigation of an alternative role for leptin during the early postnatal period has been prompted by observations that ob/ob mice show impaired central nervous system development, including reduced myelination and structural abnormalities within the hypothalamus (Bereiter & Jeanrenaud, Reference Bereiter and Jeanrenaud1979, Reference Bereiter and Jeanrenaud1980; Sena et al. Reference Sena, Sarlieve and Rebel1985). In addition, these mice fail to undergo normal reproductive maturation (Chehab et al. Reference Chehab, Lim and Lu1996), suggesting widespread impairments in the development of neuroendocrine systems. With this in mind, Ahima et al. (Reference Ahima, Prabakaran and Flier1998) sought to establish the ontogeny of leptin secretion during the early postnatal period in mice, and were able to show that there is a peak in circulating leptin levels during the early postnatal period, and that this rise does not correlate with body weight or body fat content.

In contrast to its effects in the adult rodent, during the early postnatal period leptin is unable to reduce food intake or increase energy expenditure (Mistry et al. Reference Mistry, Swick and Romsos1999; Ahima & Flier, Reference Ahima and Flier2000). Leptin receptor expression is present in the rodent ARC even during the embryonic period (Matsuda et al. Reference Matsuda, Yokota, Tsuruo, Murakami, Ishimura, Shima and Kuroda1999; Udagawa et al. Reference Udagawa, Hatta, Naora and Otani2000) and functional indications of leptin receptor signalling pathways are evident following exogenous leptin administration during the early postnatal period (Proulx et al. Reference Proulx, Richard and Walker2002; Bouret et al. Reference Bouret, Draper and Simerly2004a). It has been proposed, therefore, that although the components of leptin signalling pathways are present and functional, they are not coupled to downstream regulatory pathways until later in development. Presumably, this state of leptin resistance is beneficial in order that neonatal animals maximise their nutritional intakes.

More recently, the work of Bouret and Simerly (Bouret et al. Reference Bouret, Draper and Simerly2004a,Reference Bouret, Draper and Simerlyb) has shown that the effects of leptin during the early postnatal period are indeed different from those in adulthood. Using the lipophilic tracer, DiI, an elegant series of studies (Bouret et al. Reference Bouret, Draper and Simerly2004a) has demonstrated that in mice ARC axonal projections to downstream nuclei are immature at birth and subsequently develop during the first two postnatal weeks, and that distinct target nuclei are innervated at different stages. Furthermore, it was demonstrated (Bouret et al. Reference Bouret, Draper and Simerly2004b) that ob/ob mice have a reduced innervation of these target nuclei, that neonatal but not adult leptin treatment can reverse these neuroanatomical defects and that leptin has a direct effect on stimulation of neurite outgrowth in hypothalamic explants from postnatal animals. Thus, these studies show distinct roles for leptin at different developmental stages (either neurotrophic during development or anorectic in adulthood; Fig. 1), and in addition show that there is a critical window of timing for the leptin-mediated development of hypothalamic energy balance circuits.

Fig. 1. Diverse roles of leptin in the regulation of energy balance. ↑, Increased; ↓, decreased.

Early-life programming of appetite

Evidence from animal models has shown that permanent changes in appetite pathways can result from alterations in nutrition during early life. Plagemann and coworkers (Plagemann et al. Reference Plagemann, Heidrich, Gotz, Rohde and Dorner1992, Reference Plagemann, Harder, Rake, Voits, Fink, Rohde and Dorner1999; Davidowa & Plagemann, Reference Davidowa and Plagemann2001; Davidowa et al. Reference Davidowa, Li and Plagemann2003) have made extensive use of a rodent model in which litter size is reduced to three animals per dam for the duration of lactation in order to enhance food availability and therefore increase food intake and body-weight gain. These small-litter offspring have been shown reproducibly (Plagemann et al. Reference Plagemann, Heidrich, Gotz, Rohde and Dorner1992) to become hyperinsulinaemic in early postnatal life, to exhibit increased weight gain, adiposity and hyperphagia and to be more prone to diabetes in later life. Subsequent work using this model (Plagemann et al. Reference Plagemann, Harder, Rake, Voits, Fink, Rohde and Dorner1999) has identified alterations in appetite regulatory pathways, at the level of the hypothalamus, as a result of this postnatal overfeeding. As well as these morphological changes, it has been shown (Davidowa & Plagemann, Reference Davidowa and Plagemann2001; Davidowa et al. Reference Davidowa, Li and Plagemann2003) that these animals are less responsive to the inhibitory effects of leptin and insulin, and thus the overall outcome is an enhanced orexigenic drive.

Subsequent to these findings, and of particular importance in terms of nutritional programming of appetite, it has been shown (Yura et al. Reference Yura, Itoh, Sagawa, Yamamoto, Masuzaki, Nakao, Kawamura, Takemura, Kakui, Ogawa and Fujii2005) that alterations in maternal and neonatal nutrition can alter the profile of the postnatal leptin surge and hypothalamic development. In mice, offspring of dams subjected to undernutrition during pregnancy exhibit a premature leptin surge, and subsequently go on to show an increased susceptibility to weight gain when weaned on to a high-fat diet. The importance of this premature spike as a causal factor in the later development of obesity has been demonstrated by artificially administering a premature leptin surge in normally-fed offspring, which results in the development of a phenotype indistinguishable from that of the undernourished offspring. Thus, it appears that the development of metabolic systems critically involves this postnatal leptin surge, and that this process can be greatly affected by early nutrition. In a rat model of maternal undernutrition it has again been shown (Vickers et al. Reference Vickers, Gluckman, Coveny, Hofman, Cutfield, Gertler, Breier and Harris2005) that offspring go on to develop increased food intake and reduced energy expenditure, but remarkably that administration of leptin for 10 d during the neonatal period can correct this defect. The mechanisms in this case remain unclear. In another recent animal model it has been shown in rats (Desai et al. Reference Desai, Gayle, Babu and Ross2005) that maternal nutrient restriction and subsequent re-feeding during the lactational period results in a rapid catch-up growth during the early postnatal period, which is followed by a sustained elevated weight gain into adulthood. These animals have a lower birth weight and reduced levels of leptin at birth, but in later life attain a higher percentage body fat and became hyperleptinaemic. Together, the results of these animal studies suggest that early leptin levels provide cues that will affect the development of hypothalamic circuitry. It has yet to be determined if alterations in the postnatal leptin profile such as these have lasting effects on hypothalamic neuropeptide levels or programming of this central energy balance circuitry, but no doubt this issue will be the focus of research to come.

In human subjects there has been relatively little research into the ontogeny of early-life leptin levels. Maternal leptin levels are related to adiposity, and breast milk levels also reflect maternal fat mass (Houseknecht et al. Reference Houseknecht, McGuire, Portocarrero, McGuire and Beerman1997). Cord blood levels of leptin tend to reflect neonatal fat mass, and low cord leptin has been associated with rapid postnatal weight gain in small-for-gestational-age infants (Ong et al. Reference Ong, Ahmed, Sherriff, Woods, Watts, Golding and Dunger1999). As addressed previously, this catch-up growth is associated with an increased risk of later obesity and metabolic disease.

Animal models

Numerous animal models have now been established to study the mechanisms through which altered early nutrition and growth can affect the subsequent development of adult disease. Impaired early growth can be induced experimentally by means of maternal protein restriction (Hales et al. Reference Hales, Desai, Ozanne and Crowther1996; Langley-Evans, Reference Langley-Evans1996; Lucas et al. Reference Lucas, Baker, Desai and Hales1996), energy restriction (Vickers et al. Reference Vickers, Breier, Cutfield, Hofman and Gluckman2000), exposure to excess glucocorticoids (Seckl, Reference Seckl2004), maternal hypoxia (Williams et al. Reference Williams, Campbell, McMillen and Davidge2005), Fe restriction (Lewis et al. Reference Lewis, Petry, Ozanne and Hales2001; Gambling et al. Reference Gambling, Dunford, Wallace, Zuur, Solanky, Kaila, Srai and McArdle2003) and intrauterine artery ligation (Frampton et al. Reference Frampton, Jonas, MacMahon and Larkins1990; Simmons et al. Reference Simmons, Templeton and Gertz2001). In each of these models offspring are growth-retarded, have a lower birth weight and are more prone to disease in adulthood. In cases in which offspring are then subjected to enhanced or excessive nutrition postnatally (by the introduction of a high-fat or highly-palatable diet) the detrimental effects of this early growth restriction are augmented (Vickers et al. Reference Vickers, Breier, Cutfield, Hofman and Gluckman2000; Ozanne et al. Reference Ozanne, Fernandez-Twinn and Hales2004a). Despite different means of inducing impaired growth, the similar phenotypic attributes (altered glucose homeostasis, increased fat deposition and development of hypertension) suggest that common pathways may be involved in the programming of metabolic disease in response to poor early growth. Although extrapolation of findings in animal models cannot be directly applied to the human situation, these models allow the investigation of mechanisms that might be of importance in the development of metabolic diseases.

Of these different models, the maternal low-protein condition is possibly the best characterised, and produces offspring with a phenotype very similar to that of the human metabolic syndrome (Desai et al. Reference Desai, Crowther, Lucas and Hales1996; Ozanne et al. Reference Ozanne, Nave, Wang, Shepherd, Prins and Smith1997; Ozanne et al. Reference Ozanne, Jensen, Tingey, Storgaard, Madsbad and Vaag2005). This model therefore represents a useful tool for the investigation of mechanisms that might underlie metabolic programming, and potentially for the determination of intervention strategies.

Within the authors' laboratory the low-protein intervention involves feeding either a standard control diet containing 200 g protein/kg or an isoenergetic diet having reduced protein (80 g/kg) to female rodents during pregnancy and/or lactation. Growth restriction during gestation gives rise to offspring having a reduced body weight, and if the low-protein diet is continued throughout lactation these animals remain permanently smaller even when weaned onto a control diet (Desai et al. Reference Desai, Crowther, Lucas and Hales1996; Ozanne et al. Reference Haynes, Sivitz, Morgan, Walsh and Mark1997, Reference Ozanne, Nave, Wang, Shepherd, Prins and Smith2005). Initial studies have shown a wide range of programmed metabolic disturbances arising from these early-life dietary manipulations. In general, offspring of low-protein-fed dams show improved glucose-handling capabilities in early adult life, but undergo a greater age-related decline in glucose tolerance, and have alterations in insulin-signalling pathways in muscle and adipose tissue (Hales et al. Reference Hales, Desai, Ozanne and Crowther1996; Ozanne et al. Reference Hales, Desai, Ozanne and Crowther1996, Reference Ozanne, Wang, Coleman and Smith1997).

In an extension of this low-protein protocol cross-over groups have been included to assess the differential effects of reduced growth at different stages of early development. The crossing of low-protein offspring to control-fed dams for the period of lactation (‘recuperated’ offspring) results in a rapid growth during this period, and conversely crossing control offspring to low-protein-fed females during lactation (postnatal low-protein animals) slows growth and permanently reduces body size (Hales et al. Reference Hales, Desai, Ozanne and Crowther1996; Ozanne et al. Reference Ozanne, Fernandez-Twinn and Hales2004a). One of the most striking findings of these studies is that rapid growth during lactation significantly reduces lifespan in male rats, whereas a slowing of growth during this period increases longevity (Hales et al. Reference Hales, Desai, Ozanne and Crowther1996). More recently, the low-protein cross-over model has been applied in mice (Ozanne et al. Reference Ozanne, Lewis, Jennings and Hales2004b), and again these findings on lifespan have been replicated. In addition, it has been shown that postnatal low-protein animals are resistant to weight gain when given access to a highly-palatable diet from weaning, suggesting changes in energy balance systems. In another animal model (Vickers et al. Reference Vickers, Breier, Cutfield, Hofman and Gluckman2000) severe maternal energy restriction (to 30% control diet) results in the development of hyperphagia, obesity and hypertension in adult offspring, and additionally that these animals are hypometabolic. Furthermore, and in agreement with the authors’ findings, these detrimental effects of impaired early growth are amplified in the presence of excessive postnatal nutrition (Vickers et al. Reference Vickers, Breier, McCarthy and Gluckman2003). Although the mechanisms underlying the phenotypic responses in this model remain to be defined, recent evidence suggests that a reduction in sensitivity to the appetite-reducing and metabolic effects of leptin may play an important role (Krechowec et al. Reference Krechowec, Vickers, Gertler and Breier2006).